Enterprise IT departments can learn from hyperscalers and cloud service providers (CSPs) that deploy infrastructure at scale. These organizations that monetize their IT assets perform complex modeling and calculations to achieve an environment designed to deliver the highest performance at the lowest cost.
To accommodate these complex models, organizations deploying servers at scale tend to measure performance and power at the rack level. That is, the amount of work a rack of servers can perform and the power envelope. And for cloud-native workloads, which tend to be lightweight and have consistent performance, work is measured in cores.
As mentioned above, all major CSPs are deploying Arm-based servers to support general purpose customer workloads. Also, given the design and performance of Ampere’s Altra Max CPUs, it’s no surprise that they are the CPU of choice for all CSPs who haven’t designed their own Arm chips.
In the next few sections, we’ll look at the evolution of the data center processor market and where Ampere offers compelling solutions that offer performance, cost, and sustainability. We also delve into rack-level measurements for power and performance and why we believe they are the right metric.
CPU Market – A Brief History
The enterprise data center has long been a space dominated by x86, with two major players driving the innovation cycle of software developers and IT professionals supporting business needs. Mission and business critical applications developed in a three-tier model were deployed at scale across enterprise data centers. As computing requirements grew, IT organizations added servers and storage.
The above operating model has served the business very well for a very long time. In addition, as virtualization technology has been adopted by enterprises, operational efficiencies have further improved as deployment and management of virtual servers reduces time to production.
As cloud computing has evolved from its early days into an essential part of the enterprise data center, cloud-native architecture has become the preferred method for designing and deploying applications for performance, mobility, and security reasons. rice field. And cloud native isn’t just about developing new applications. Many organizations are refactoring and redesigning their traditionally designed mission and business-critical applications as part of their digital transformation efforts.
Cloud-native workloads are designed to be lightweight and mobile. The concept is that the application runs only when needed and is split into various microservices that can be shared with other applications. And these services move from server to server, from on-premises to off-premises, and from cloud to cloud.
I’m not writing this to educate readers who already understand cloud native concepts, but to map these needs to the evolution of CPUs in the cloud. Many of these lightweight services continuously spin up and down, benefiting from servers with CPUs with many single-threaded cores to support these performance characteristics. And this is essentially an Arm CPU. When AWS first introduced his homegrown Graviton CPUs, they were quickly successful in supporting these proliferating general-purpose cloud workloads for businesses of all sizes and types. And for analysts (and the industry), Graviton’s success proved Arm’s viability in cloud (and hyperscale) data centers. This is because it delivers unparalleled levels of compute density and reduces the power envelope, resulting in real and measurable cost savings. .
Please enter the amperage
While AWS worked on internal development of Graviton, a few former Intel executives and semiconductor veterans quietly worked on designing and delivering a commercial Arm CPU that could support everyone else. Please enter Based on Arm’s Neoverse Architecture, Altra Max is a CPU that combines the performance, reliability, and consistency of x86 with the power savings associated with Arm.
Ampere positions Altra Max as the first cloud-native CPU on the market. I like this positioning because it describes exactly where the CPU is running in CSP and hyperscale environments. These are organizations that deploy hundreds of thousands of servers and tend to measure performance and power at the rack level.
Ampere’s positioning resonates with CSPs. Following AWS’ success with Graviton, the remaining major cloud providers have embraced Altra Max, offering Arm-based instances as an economical alternative to support cloud-native workloads. Oracle (an investor in Ampere) first launched Oracle Ampere A1 Compute instances and made them available in all regions. Microsoft Azure followed suit with his three instance types in the Dpsv5, DPlsv5 and Epsv5 series (very original naming). And Google Cloud Platform (GCP) has lagged behind with the launch of the Tau T2A instance.
There are two perspectives on Ampere adoption in the cloud. First, AWS put competitive pressure on the market to launch Arm-based instances. Ampere was a great story. Second, other his CSPs saw the popularity of his Graviton and the internal cost savings achieved by deploying Arm at scale, making Ampere a competitive (and attractive) commercial provided the product. Knowing the CSP market, I invest all my money in the latter as it has tangible and measurable advantages in terms of cost and manageability. And these costs are passed on to our customer base.
A deep dive into ampere efficiency with rack talk
Ampere’s value proposition is in terms of compute density and power efficiency per rack. In other words, how many compute cores can fit in a 42 rack unit (RU) form factor and how much power does that data center rack require? For performance workloads, this rack level measurement is the best measurement. This allows data center architects to quickly estimate compute capacity for large-scale needs. It can also measure power consumption and power efficiency very quickly and accurately.
It is also important to measure both performance and power at the rack level. It explains how to achieve maximum efficiency, a challenge for nearly every data center architect. As a data center architect, if I could support cloud-native workloads on 3,300 racks instead of 10,000, it would save a lot of data center floor space (and power, management, etc.). The first obvious advantage is his 1/3 cost reduction in the space required for the server farm. The second obvious benefit is lower monthly electricity bills. This is especially important in today’s skyrocketing energy costs. And the third is the less noticeable (but significant) cost savings achieved by deploying, monitoring, and managing fewer servers.
These savings are significant. And this is exactly what Ampere claims.
It’s important to note that the above compares Ampere to a competitor’s N-1 configuration. AMD’s 4th Gen processors support up to 96 cores (soon to be 128 cores) per CPU. Similarly, Intel’s 4th Gen Xeon processors support up to 56 cores. That being said, 3rd generations of processors hit the market.
Also note that Moor Insights & Strategy did not run the tests and did not perform a detailed methodological analysis, but the tests passed the first pass analysis.
Ampere’s claim is impressive. Almost too good to be true. But, as I said before, the numbers do add up. And the company has put a lot of noise behind campaigns pointing out these benefits to the market.
Don’t forget sustainability
Another area Ampere is focusing on is its apparent commitment to sustainability. His 2.8x reduction in electricity consumption adds up massively, as reducing carbon footprint is a top priority for virtually all companies (and indeed all his CSPs).
There are two perspectives on sustainability: the idealistic ‘save mother earth’ perspective and the realistic ‘cost cutting’ perspective. Each one is legitimate and important to the enterprise. Again, Ampere makes a very bold statement. Cloud data centers supporting cloud-native workloads with Ampere can go a long way toward achieving your sustainability goals. In addition, power consumption alone can result in astronomical cost savings.
Conclusion – Is this too good to be true?
Good performance combined with a reduced carbon footprint and significant cost savings sounds too good to be true. And I’m sure x86 players will respond to Counter Ampere’s claim. Isn’t that how these campaigns work?
Ampere has a point. We no longer live in a world of data centers where x86 is the only option, and especially in the cloud, there is no longer an “yes but” mindset to Arm deployments. The ecosystem no longer strongly distinguishes between architectures, and indeed the cloud-native software ecosystem has enthusiastically supported Arm as a supporting platform. Yes, rack power is a good metric.
Ampere has established itself in the data center and is arguably the leading chipmaker for the cloud. And this competitive pressure is good for everyone, especially customers.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.
Moor Insights & Strategy does or has provided paid services to technology companies, including all research and technology industry analyst firms. These services include research, analysis, advice, consulting, benchmarking, winning matchmaking, video and speaker sponsorships. The company includes 8×8, Accenture, A10 Networks, Advanced Micro Devices, Amazon, Amazon Web Services, Ambient Scientific, Ampere Computing, Anuta Networks, Applied Brain Research, Applied Micro, Apstra, Arm, Aruba Networks (now HPE), Atom Computing, AT&T, Aura, Automation Anywhere, AWS, A-10 Strategies, Bitfusion, Blaize, Box, Broadcom, C3.AI, Calix, Cadence Systems, Campfire, Cisco Systems, Clear Software, Cloudera , Clumio, Cohesity, Cognitive Systems, CompuCom, Cradlepoint, CyberArk, Dell, Dell EMC, Dell Technologies, Diablo Technologies, Dialogue Group, Digital Optics, Dreamium Labs, D-Wave, Echelon, Ericsson, Extreme Networks, Five9, Flex, Foundries.io, Foxconn, Frame (now VMware), Fujitsu, Gen Z Consortium, Glue Networks, GlobalFoundries, Revolve (now Google), Google Cloud, Graphcore, Groq, Hiregenics, Hotwire Global, HP Inc., Hewlett Packard Enterprise, Honeywell, Huawei, HYCU, IBM, Infinidat, Infoblox, Infosys, Inseego, IonQ, IonVR, Inseego, Infosys, Infoot, Intel, Interdigital, Jabil Circuit, Juniper Networks, Keysight, Konica Minolta, LatticeS semiconductor, Lenovo, Linux Foundation, Lightbits Labs, LogicMonitor, LoRa Alliance, Luminar, MapBox, Marvell Technology, Mavenir, Marseille Inc, Mayfair Equity, Meraki (Cisco), Merck KGaA, Mesophhere, Micron Technology, Microsoft, MiTEL, Mojo Networks, MongoDB, Multefire Alliance, National Instruments, Neat, NetApp, Nightwatch, NOKIA, Nortek, Novumind, NVIDIA, Nutanix, Nuvia (now Qualcomm), NXP, onsemi, ONUG, OpenStack Foundation, Oracle, Palo Alto Networks, Panasas, Peraso, Pexip , Pixelworks, Plume Design, PlusAI, Poly (formerly Plantronics), Portworx, Pure Storage, Qualcomm, Quantinuum, Rackspace, Rambus, Rayvolt E-Bikes, Red Hat, Renesas, Residio, Samsung Electronics, Samsung Semi, SAP, SAS, Scale Computing, Schneider Electric, SiFive, Silver Peak (now Aruba-HPE), SkyWorks, SONY Optical Storage, Splunk, Springpath (now Cisco), Spirent, Splunk, Sprint (now T-Mobile), Stratus Technologies, Symantec, Synaptics, Syniverse, Synopsys, Tanium, Telesign, TE Connectivity, TensTorrent, Tobii Technology, Teradata, T-Mobile, Treasure Data, Twitter, Unity Technologies, UiPath, Verizon Communications, VAST Data, Ventana Micro Systems, Vidyo, VMware, Wave Computing, Wellsmith, Xilinx, Zayo, Zebra, Zeda, Zendesk, Zoho, Zoom, and Zscaler. Moor Insights & Strategy Founder, CEO and Chief His Analyst Patrick Moorhead is an investor in dMY Technology Group Inc. VI, Fivestone Partners, Frore Systems, Groq, MemryX, Movandi, and Ventana Micro.